Is it canine DUSN?

Abstract

The term retinopathy has been used to group several heterogeneous retinal abnormalities that are clearly acquired or are suspected/proposed to be inherited. Some share characteristic focal/multifocal hyperreflective tapetal lesions with a dark center, and areas of non-tapetal depigmentation suggestive of patchy or diffuse outer retinal atrophy. Progression is variable, and some develop unilateral or bilateral fundus changes resembling the clearly inherited form of retinal degeneration referred to as PRA. In this Commentary and Review we discuss the role of ocular larva migrans resulting in the entity we refer to as canine DUSN, and suggest that it may be responsible for some of the retinal findings grouped under the retinopathy rubric that share this characteristic fundus lesion.

Introduction

Over the past 20+ years, an explosion of genetic/genomic information has enabled studies identifying genetic traits and diseases in dogs; the lack of genomic resources before 1990 prevented such studies. However, the first linkage map of the dog genome published in 1997¹ led to the rapid identification of the first autosomal retinal disorder of dogs, progressive rod-cone degeneration (prcd)². Now, high density SNP chips as well as targeted, exome, whole genome or RNA sequencing, etc. are possible due to the sequencing of the canine genome^{3, 4}. Leading the way in developing the needed tools and resources has been the study of inherited retinal diseases, and there are now more identified genes/mutations in canine retina than for any other organ system^{5, 6}.

So what evidence is required to confirm that a presumed inherited retinal disease is indeed inherited? Careful and detailed phenotyping of affected dogs ,and delineation of the natural history of disease to determine consistency of onset, progression and severity in a cohort of study animals are essential. Most importantly, genetic studies are foundational in establishing the inherited basis, requiring either prospective informative matings, or detailed retrospective pedigree analysis. Sadly, this is not always the case in veterinary ophthalmology, and publications often attest to the putative heritability of a disorder based on “… a review of the pedigree suggests/indicates/confirms that the disease is inherited as a simple recessive/dominant/incompletely dominant etc. trait…” without providing the pedigrees or any details to support such a conclusion.

The term retinopathy has been used extensively in veterinary ophthalmology with increasing frequency but lack of specificity. Some diseases are clearly inherited and have a characteristic phenotype and known molecular defect^{7, 8}. Others are the result of viral diseases, e.g. distemper retinopathy; however, many dogs receive a retinopathy diagnosis based on fundus lesions that, without any evidence, are suspected to be inherited.

This commentary and review will be limited to a group of ‘retinopathies’ that have a surprising commonality in clinical disease characteristics, including similar and often asymmetrical ophthalmoscopic appearance, and a somewhat unpredictable disease progression or clinical outcome. Overall, these dogs have focal to multifocal areas of tapetal hyperreflectivity, often with a central dark area referred to as pigmented, and areas of non-tapetal depigmentation suggestive of patchy or diffuse outer retinal atrophy. Unilateral lesions can remain unilateral, become bilateral, and/or expand. Enlarging progressive fundus changes may suggest generalized progressive retinal atrophy (PRA) with vascular attenuation, although telltale lesions recognized at earlier examination timepoints may still be present. Some of these retinopathies are either clearly acquired^9–11, proposed as inherited¹² or presumed inherited^{13, 14} (Supplemental Table 1).

One consistent finding regardless of whether the retinal abnormalities are considered inherited, sporadic or acquired is the preponderance of affected males. The skewed male to female ratio (M/F) ranges from 4/1 to 11/1 (Supplemental Table 1). This is clear when examining the results of studies of ocular larva migrans (OLM) multifocal retinitis in New Zealand sheep dogs⁹ or Border collies¹¹, Borzoi retinopathy^{13, 14}, sporadic retinopathy in German shepherds¹⁰ and retinal degeneration in Greyhounds¹⁵. The skewed male to female ratio in Border collies has been explained by the disorder being X-linked¹². While most explanations are plausible and reasonable perhaps it is time to think outside the box and reassess how we diagnose and categorize these retinopathies.

It’s Canine DUSN, Doofus!

Since the early 1950s, parasitic granulomas resulting from nematode larvae (Toxocara spp.), initially referred to as visceral larva migrans (VLM) and subsequently OLM, have been considered an important differential for retinoblastoma, a highly malignant childhood cancer¹⁶. Much smaller lesions confirmed by histopathology also were found in young Beagle dogs during routine ophthalmic screening^{17, 18}. More recently, I (GDA) have been struck by how often I examine dogs mainly of mid-/large-size breeds that show fundus findings broadly characterized by focal to multifocal hyperreflective lesions with/without patchy areas of depigmentation in the non-tapetal region. These can be unilateral, and the fellow eye normal, or can have more extensive bilateral retinal thinning and vascular attenuation. My general diagnosis was ‘non-progressive healed focal/multifocal retinal inflammation’. However, some did progress in spite of my unfounded reassurances. Other dogs had similar lesions with additional single or multiple granulomas, either sub- or pre-retinal. On the most severe end of the spectrum, granulomas with retinal traction bands were the predominant fundus finding. Only rarely were subretinal nematodes found on clinical examination, but these were not documented photographically. Figures 1 and 2 illustrate some of the fundus abnormalities identified.

Figure 1.

Focal/multifocal retinal lesions. (A, B) Unilateral focal tapetal hyperreflectivity with darker and diffuse center in a (A) 5.3 year male Borzoi, left eye, and (B) an adult male Australian shepherd. (C_1,2) 6.5 year male Greyhound, left eye. Unilateral multifocal hyperreflective lesions in tapetal region with a small dark center. (D) 7.5 year old male Border collie, left eye. There are multiple, large areas of tapetal hyperreflectivity only in the left eye. Within these areas are smaller single or multiple dark areas, often surrounded by a green halo. The optic disc is pale with minimal surface vessels. (E_1–3 and F_1,2) Male Borzoi at 3 (E_1–3, right eye) and 4 (F_1,2, left eye) years of age. The right eye has focal tapetal and non-tapetal lesions; the non-tapetal lesion has a central dark/gray scar. (F_1,2) The left eye, which was normal at 3 years of age, now has multiple areas of tapetal hyperreflectivity surrounding the optic disc (F₁). The more peripheral tapetal region shows more diffuse hyperreflectivity and slight vascular attenuation (F₂). Note that the light intensity used for the F₂ photograph was reduced by 1 log unit.

Figure 2.

Focal/multifocal retinal lesions with pre-/sub-retinal granulomas and vitreous traction bands. (A) young adult Siberian husky male, right eye. Unilateral focal area of tapetal hyperreflectivity with 2 central dark foci. Within ~ 1 disc diameter is a subretinal granuloma with apparently 2 much smaller subretinal satellite lesions. (B) 7 year old male Borzoi, left eye. Multifocal areas of tapetal hyperreflectivity with dark center. Overlying the superior central retinal vessels is an area of vitreous condensation which blurs the retinal vessel detail. (C_1,2) adult Australian shepherd male, right eye. A unilateral, faint focal hyperreflective lesion (C₁, arrow) with a dark center is located in the temporal quadrant near a well demarcated subretinal granuloma. There is vitreal condensation overlying and surrounding the optic disc which shows mild swelling. (D) 6.7 year old Borzoi male, left eye. Focal hyperreflective area with a dark center (arrow) is located adjacent to a retinal arteriole. More peripherally is a subretinal granuloma with a focal hemorrhagic area on its inferior border. (E_1,2) left and (F_1–3) right eyes of a 3 year old male Borzoi. The left eye has a large area of tapetal hyperreflectivity temporal to the disc (E₁) and the non-tapetal region has irregular patches of depigmentation and a single circular depigmented area with a central white scar (arrow, E₂). (F₁) The right eye has a temporal juxtapapillary subretinal granuloma (white arrowhead), and a nasal hyperreflective lesion, approximately 1 disc diameter, with a dark core in the periphery (black arrow). There is mild swelling of the optic disc. (F_1–3) Originating from the inferior surface of the optic disc is an elongated vitreous traction band that divides as it extends into the inferior periphery. (G_1–3) right and (H_1,2) left eyes of a 3 year old male Alaskan malamute. (G_1–3) The right eye has a 1.5 disc diameter sized hyperreflective lesion with 2 dark foci underlying the superior retinal venule, and an area of vitreal condensation supero-temporally (G₁, arrow). (G_2,3) The non-tapetal zone has 2 focal retinal granulomas that are surrounded and joined by a vitreal traction band. (H_1,2) The left non-tapetal zone has multiple granulomas located subretinally (arrowhead) or on the retinal surface (arrow).

Because of my concern that these retinal findings were increasing in frequency within the limited population of dogs in my clinical practice, I contacted Kevin R. Kazacos at Purdue University to discuss the possibility that these fundus lesions represented OLM from either Toxocara canis, Baylisascaris procyonis or another nematode. Kevin has published extensively on OLM in animals and humans, and is an authority in the field (selected publications include^{9, 19–27}). After explaining my findings and their possible interpretation, his first question was: “Are you familiar with DUSN?”. A long silence followed my reply of “no…..” during which I am sure that the thought ‘It could be DUSN, Doofus!’ went through his mind. Luckily, he was too polite to verbalize it! I was surprised by my apparent lack of familiarity with Diffuse Unilateral Subacute Neuroretinitis (DUSN) as I had read several articles on the topic in both man and experimental animals, and had annotated reprints in my files. However, it never seemed to rise to the level of awareness in my clinical assessment as I hadn’t considered that it might occur in dogs.

Diffuse Unilateral Subacute Neuroretinitis (DUSN)

In the late 1970’s, Gass and Scelfo reported a “… peculiar uniocular syndrome affecting children and young adults characterized by early loss of central vision, vitreous inflammation, papillitis and recurrent crops of evanescent, gray-white lesions affecting the outer retina and pigment epithelium and later by progressive loss of visual field, optic atrophy, narrowing of the major retinal vessels, diffuse as well as focal atrophic pigment epithelial changes throughout the fundus …”²⁸. The condition was termed DUSN, and subsequently additional patients were described with similar clinical signs and subretinal nematodes, i.e. OLM²⁹. DUSN appeared to be caused by two different-sized nematode larvae; one was 400–1000μm long, and the second 1500–2000 μm, and their length seemed to be specific for the patient’s geographic region²⁸. To date, many cases of DUSN have been reported worldwide, and is especially prevalent in South America, where the small worm is endemic^30–34. Even though DUSN is considered a uniocular disease, bilateral cases also occur^{30, 32}, likely attributed to the worm load. Indeed, the increased worm load in dogs, which have ample exposure to canine, raccoon or other feces with roundworm eggs and the increased biological propensity for somatic migration of larvae likely contributes to the more frequent bilaterality of disease in dogs. As well, the continuum of fundus abnormalities ranging from focal to multifocal hyperreflective lesions with/without non-tapetal patchy areas of depigmentation to more advanced retinal degeneration and sub-/pre-retinal granulomas likely reflects levels and frequency of exposures.

Gass’ 1983 publication stressed the importance of locating the worm in patients with a DUSN diagnosis²⁸ because early treatment by laser photocoagulation is recommended³⁵. However, not all patients have a visible nematode on the first visit, and multiple re-examinations often are necessary using biomicroscopy and a fundus contact lens, 3-mirror Goldmann lens or Volk 78D lens^33–35. In veterinary ophthalmology, the routine methods used for clinical examinations may not offer the magnification necessary to confirm the presence of a small sub/intra retinal nematode.

DUSN and Larva Migrans (LM)

Based on the presence of helminth larvae, DUSN is a form of OLM, and OLM can have numerous causes in man and animals^36–39. What sets all forms of LM apart from typical parasitic infections is that they are related to somatic migration of helminth larvae, usually occurring in paratenic hosts as part of their normal life cycle⁴⁰. Many incorrectly view OLM and other kinds of LM as “aberrant migration” of larvae, but in reality, this isn’t true. Somatic migration with dissemination of larvae into various organs and tissues is normal behavior required for completion of their life cycle, and OLM represents an accidental entrance into the eye.

Causes of LM, including OLM, are typically helminth parasites of carnivores, ranging from dogs and cats to wild mammals and even reptiles^{36, 38, 39}. Carnivore helminth larvae have evolved to infect, migrate in, and eventually become encapsulated in organs and tissues of paratenic and intermediate hosts, surviving for extended periods of time in a state of arrested development. These hosts, destined to become the carnivore’s prey, commonly include various rodents, rabbits, birds, etc. Following ingestion by the carnivore definitive host, the larvae are released, activated, and complete their development to adults, usually in the GI tract.

Two of the most common potential causes of LM are Toxocara canis/T. cati of dogs and cats, and Baylisascaris procyonis and related species in wild carnivores (raccoons, etc.)^{36, 39}. T. canis and B. procyonis are most often implicated in cases of VLM, neural LM (NLM), OLM and DUSN in humans and animals due to their commonality as pet and wild animal infections, leading to widespread environmental contamination with resistant eggs and subsequent exposure and infection^{38, 39}. Depending on the levels of infection and the particular organs/tissues involved, cases will range from overtly clinical to subclinical/asymptomatic³⁹.

Helminths in DUSN

The cause of DUSN initially remained enigmatic until the fortuitous observation of nematode larvae in the eyes^28,29. The small nematode variant (400–1000μm) most likely involved classic Toxocara OLM, but this conclusion was originally discounted by medical ophthalmologists because the patients were seronegative for toxocariasis despite waning titers, time lags in testing and sequestration of antibodies in vitreous fluid^{38, 39}. The larger nematode (1500–2000μm) was proposed to be B. procyonis by one of us (KRK) based upon its commonality and exact size match with the larvae in large-nematode variant DUSN cases^22–24. By that time, Baylisascaris was a well-known cause of LM in animals and humans, with numerous cases of NLM and mounting evidence of its involvement in OLM. However, its role in DUSN was originally dismissed out of hand, as patients did not have concomitant or preceding NLM/CNS disease, nor were considered to have raccoon exposure because they lived in cities⁴¹. Both assertions were soundly debunked^{26, 42}; experimental studies producing OLM-DUSN with B. procyonis in two species of non-human primates and various rodents^{23, 24, 39} confirmed its role in the disease.

On the basis of the association between the intraocular migration of nematodes in man and DUSN, it is now appropriate to consider the possible role of nematodes in one form of retinopathy we refer to as canine DUSN. To set the stage, it is important to first review two seminal studies of OLM in dogs.

The 1987 study by Hughes and associates was the first large-scale analysis of multifocal retinitis in New Zealand Heading and Huntaway sheep dogs⁹. Of the 1448 dogs examined (m=990; f=458) 565 were affected with multifocal retinitis that in its most severe form consisted of vitreal haze with track-like granulomatous inflammation; ~half were affected bilaterally. Noteworthy is that 50% of the male dogs examined had retinal lesions versus 15% of the females (Supplemental Table 1).

Eyes were available for histopathology from many of the dogs. Those grouped into Category I disease were young dogs with perivascular infiltrates in the inner retina and track-like granulomatous inflammation, and ~half had Toxocara larvae in subretinal tracks or granulomas; larvae were not identified in eyes with more advanced retinal lesions⁹. Interestingly, subretinal larvae were seen primarily in cases equivalent to early-stage DUSN in man³³.

Studies by Johnson and associates in Border collies provided strong evidence that OLM and DUSN occur in dogs, and were the first to make this association¹¹. Both ophthalmoscopic and histopathogical changes were similar to the lesions seen in the New Zealand study. Three of the 4 affected dogs were fecal positive for Toxocara eggs, and their tissues (including eyes) contained eosinophilic granulomata consistent with encapsulated Toxocara larvae; an intact third-stage Toxocara larva was digested from part of the inferior retina of one.

To ascertain that environment and not genetics was responsible for the disease, Hughes examined 195 non-sheep dogs that were kept as pets in either a rural or urban environment; the rural environment being comparable to that of the working sheep dogs⁹. Approximately 43% of the 70 rural pet dogs had multifocal retinal lesions similar to the ones found in the sheep dogs; of the 30 affected dogs, only 2 were females. In contrast, only 6% of 125 urban pet dogs had retinal lesions, and all were males. Interestingly, all four affected young dogs reported by Johnson and associates were also males, and the sire of the litter had a single hyperreflective focal lesion with a dark center¹¹ (Johnson and Whiteley, personal communication to GDA 2020) (Supplemental Table 1).

In addition to environment, diet may have an influence. Humans can contract Toxocara VLM by eating various raw organs and meats from animals that harbor migrating larvae and that many animal species are potentially infected^43–45. Hughes and associates noted that sheep dogs and rural pet dogs in New Zealand were often fed a diet heavily consisting of uncooked but frozen mutton. Similarly, in the case of the Border collies, the dogs were fed pork scraps¹¹. T. canis larvae can remain viable for several weeks or more in frozen carcasses⁴⁶. In recent years, there has been an uptick in “natural” or “raw” diets for dogs, consisting of raw meats of various types, which could expose dogs to infection if the meat came from animals exposed to T. canis egg contamination.

From these two studies, it is clear that canine DUSN occurs, and that environment/husbandry/diet determine the exposure to parasite eggs and migrating larvae and therefore disease severity^{9, 11}. It is also clear that within the study population the lesions are not uniformly distributed within and between eyes, or between dogs. Finally, it is essential to recognize that even when there is presumably equal exposure to parasite eggs or larvae, the frequency of disease between males and females is not the same, and females are either ‘protected from’ or males are ‘favored to’ developing retinal lesions (see below).

Selected Retinopathy Disorders (Supplemental Table 1)

Borzoi retinopathy

An abstract in the 1977 Proceedings of the ACVO reported a retinal degeneration in the Borzoi breed¹³. The brief report included a limited pedigree from one kennel suggesting “an autosomal recessive mode of transmission”. The report indicated that progeny “from affected to affected matings are currently being evaluated and a variety of test matings are underway”, but results were never published. However, sometime later the term Borzoi retinopathy entered the lexicon of veterinary ophthalmology, and has been used since. The abstract summarized the findings in 180 dogs of which 20/91 males and 5/89 females were affected. The age distribution at diagnosis was broad, 6m to 9 years, with 3 of the severely affected dogs less than 4 years old. The reported lesions were multiple foci of retinal degeneration which increased in size and number, coalesced and progressed to diffuse retinal atrophy. Eight cases were unilateral at first examination, and early lesions in some were in the tapetal periphery.

The first actual publication on the disease was a study by Chaudieu describing the clinical evaluation of 160 dogs with 8/63 males and 1/97 females affected¹⁴. Lesions (5 unilateral and 4 bilateral) were found in animals 1.5–6 years old, and all located in the tapetal zone with either focal hyperreflectivity with or without a pigmented center, and no progression to advanced degeneration. Based on detailed pedigree analysis, autosomal recessive inheritance was proposed. A similar lack of progression was reported in a more recent study in 103 dogs, but pedigree analysis was unable to confirm that the condition is inherited⁴⁷. These three studies grouped under the Borzoi retinopathy rubric do not show the same retinal changes, age of onset or progression suggesting that they may be causally distinct, and have only in common a retinal finding found in Borzois.

Sporadic retinopathy

Retinopathy in German shepherd police dogs consisted of focal/multifocal areas of tapetal hyperreflectivity with a pigmented center, and a mean age of diagnosis of 6.8 years¹⁰. Lesions were generally unilateral on the first exam, and could become bilateral and larger without progression to blindness or vision deficits impairing work. Of 96 dogs examined, 86 were males and 10 females (personal communication Ireneusz Balicki, 7/6/20 and¹⁰). However, only male dogs were affected (n=10). The authors concluded that the retinal abnormalities are not hereditary or secondary to systemic disease, and proposed micro hemorrhages resulting from stress/exertion as potential causes for the retinal finding citing a 2004 ACVO abstract in support of this interpretation; a paper based on that abstract was never published. Similar fundus abnormalities were found in racing Greyhounds in Australia with only male dogs being affected¹⁵ (Supplemental Table 1).

X-linked progressive retinal atrophy (XLPRA) in Border collies. A primary retinal disease?

Vilboux and associates reported a form of XLPRA in Border collies¹². Of 487 dogs examined by the same ophthalmologist, 54/241 males and 6/246 females were affected with asymmetric focal or multifocal hyperreflective lesions of the tapetum resulting in thinning of the retina. Generalized hyperreflectivity coexisted either with hyperreflective coalescent foci in the tapetum, or with diffuse tapetal hyperreflectivity and depigmentation/hyperpigmentation in the nontapetum; the vessels became thinner before disappearing and the optic disc became grayish. Lesions were unilateral or bilateral at early and advanced stages of the disease, and some unilateral abnormalities on the first examination became bilateral. The ERG recorded from dogs with isolated central or peripheral lesions was normal, but became abnormal in terms of waveforms and amplitudes with advanced disease.

Detailed study of an 80 dog pedigree and segregation analysis concluded the retinal abnormality is inherited as a fully penetrant X-linked recessive trait¹². More recently, 178 related Border collies, 36 affected, were genotyped on the Illumina HD 170K SNP array, and whole genome sequencing carried out on 5 related dogs. Although 78 single nucleotide variants (SNVs) were found, these were not in coding regions of X-chromosome genes⁴⁸.

In our opinion, the retinal abnormalities reported for the Border collie fail to meet the criteria for an inherited retinal defect, and likely represents a non-inherited retinopathy. In a classic slide by Dr. Donald Paterson, University of Pennsylvania, and copied without attribution prior to the 12^th Blue Book edition, he set out 6 criteria one should use to suspect that a disorder is inherited in a breed. The Border collie retinal disease meets 3 of these criteria: 1) when the frequency is greater than in other breeds; 2) when the frequency increases in a given breed as a whole and 3) when the frequency is greater in related dogs within a breed), but not others: 4)-the disease does not have a characteristic appearance and location; focal and multifocal hyperreflective lesions are not the same, but likely represent multiple insults to the same tissue; 5)-while the disease may have a characteristic age of onset in general terms, it lacks a predictable course of progression, i.e. no predictable stages for development and time for each stage to development; 6) the disease does not look identical to an entity which has been proven to be inherited in another breed. In fact, it is entirely the opposite, as it looks just like some acquired retinopathies that are sporadic in other breeds.

Given the very strong genetic support for XLPRA in Border collies how can we reconcile the clinical and genetic data? To do this, we need to look at the gender distribution for the other retinopathies and summarized in Supplemental Table 1. In all, there is an overwhelming preponderance of affected males vs females. Thinking outside the box, we propose that the retinal findings are acquired, but that there is/are genetic or gender modifier(s) in the X-chromosome that either favor the development of disease in males, or protects the females (see below).

Focal/multifocal tapetal hyperreflectivity with central ‘pigment’ clump; the common retinal finding shared among diverse retinopathies

The common clinical descriptor for many of the retinopathy publications summarized here is the presence of focal to multifocal areas of tapetal hyperreflectivity with a central ‘pigment’ clump early in the disease (Figure 1A–D, 2A, B, BF1, BG1). The assumption is that the dark center represents focal proliferation of the RPE, although confirming histopathology is lacking. An alternative is that the central focal area represents focal tapetal/RPE damage from a local insult. Similar focal/multifocal lesions are associated with the resolution of any retinal inflammatory process; among these, canine distemper retinitis was, at one time, the most common cause of such lesions¹⁷. In my (GDA) experience, similar lesions occur in dogs with embolic phenomena where the emboli lodge in the arterioles that course perpendicular through the tapetum and connect the mid-sized choroidal arterioles to the choriocapillaris (Figure 3A, B). Initially these infarcts appear as a bull’s eye with a ring of subretinal edema surrounding a central lesion that can have a central hemorrhage or ‘pigment’. Once edema subsides, the retina flattens, atrophies and a hyperreflective ring surrounds a central dark area.

Figure 3.

Anatomic considerations explaining possible routes of entry of nematodes into the subretinal space. (A) Low magnification image of the retinal tapetal region and choroid. The large outer choroidal vessels progressively become smaller, and fine arterioles (arrows) penetrate through spaces in the cellular tapetum (180x). (B) Higher magnification of the same area showing details of two arterioles (yellow arrows in A and B) that branch and form the choriocapillaris layer between the basal retinal pigment epithelium (RPE) and innermost tapetal cells (560x). (C, D) Non-tapetal region showing areas where there is limited choriocapillaris-basal RPE interactions (C, 740x) and other areas where there is extensive choriocapillaris layer (D, *; 1150x). (E) Tapetal region where there is a paucity of arterioles that penetrate through the tapetal layer (TL; 800x). (F, G) Electron micrograph of the same general area showing the association of the RPE with the tapetum lucidum (TL), and the clearly demarcated Bruch’s membrane (BM). In this intercapillary zone, the choriocapillaris layer is absent (F=3500x; G=20,000x). Such areas likely would serve as an impenetrable barrier to nematodes and prevent their access to the subretinal space.

Is it canine DUSN?

We propose that intraretinal/subretinal migrating nematode larvae should be considered as causative/potentially causative for some of the retinal diseases grouped under the retinopathy rubric, and characterized by similar clinical findings. Based on the close parallel between the ocular findings in human patients with DUSN and dogs with these characteristic findings, particularly when advanced, it is important that OLM be considered a potential cause, and certainly should be included in the list of differential diagnoses (Table 1). The frequency and severity of disease in human patients occurs in populations with varied exposure to areas contaminated with dog, raccoon, and other animal feces. Unlike humans, many dogs are coprophagic and eat various detritus and dead animals, resulting in higher levels of infection and LM. In addition, it is more common for dogs (versus humans) to have access to and consume raw meat as part of their diet. Based on the load of parasite eggs and larvae ingested, this would account for the more diffuse and bilateral retinal findings in dogs. If canine DUSN is the diagnosis, there are several points that need consideration. In contrast to the focal/multifocal hyperreflective lesions described above, there is little question about a canine DUSN diagnosis when accompanying posterior segment granulomas (Figure 2). On detailed fundus examination, it is usually possible to locate the areas of intraocular penetration, often in close proximity to subretinal/preretinal granulomas. In more severe cases, vitreo-retinal traction bands are present.

Table1.

Comparison of clinical findings in patients with DUSN diagnosis and the proposed canine clinical homologue

	DUSN*	canine DUSN
Severe loss of peripheral and central vision	+	Only with advanced retinal degeneration
Multiple gray-white lesions at the level of the outer sensory retina and pigment epithelium	+	±
Patchy areas of RPE atrophy evolved into progressive widespread irregular areas of focal or diffuse RPE depigmentation	+	+
Narrowing of the retinal vessels	+	In advanced cases or with more generalized retinal degeneration
Optic disc swelling early; optic atrophy late	+	+
Vitreous cells; vitritis; vitreo-retinal traction bands between disc and granuloma	+	+
Subretinal nematode	+	+, but rarely identified, and not documented photographically
Subretinal granulomas	+	+
Fluorescein angiography early active	normal gray white, deep, retinal lesions appeared hypofluorescent during the early stages of angiography and they stained during the later stages	Borzoi: early stages normal, and in one dog neovascularization suggestive of choroidal inflammation14 Borzoi: acute lesions transiently hyperfluorescent; no progression47 Border collie (XLPRA?): window defect and masking by focal hyperreflective lesion in choroidal phase; in laminar venous and late venous phase, focal leakage in RPE and diffusion of fluorescein at lesion periphery12
Electroretinogram early late	normal abnormal; rod/cone function decreased	Borzoi: 4 dogs have low amplitude ERG, but only focal/multifocal disease14 Borzoi: early disease. normal47 Border collie (XLPRA?): normal for isolated central or peripheral lesions; abnormal and decreased amplitudes in advanced disease12 German shepherd: impaired cone function with lower b-wave amplitude and flicker, prolonged implicit time in mixed rod-cone, photopic and flicker recordings10
Unilateral or bilateral	Unilateral; 2 bilateral cases: 1 of 36 patients29, and 1 patient32	Unilateral or bilateral
Gender distribution (M/F ratio)	22/1529 45/3330 12/628	Overwhelming male preponderance-see Supplemental Table 1.

^*References for clinical characteristics of DUSN are:^{16, 28−30, 32, 35}

Gender disparity in retinopathy/canine DUSN

There is a striking gender disparity in cases grouped under the retinopathy rubric, and in the disease we now propose as canine DUSN. The cause(s) for this male predominant disease are unknown. Obviously, if inherited as a fully-penetrant X-linked recessive trait as proposed for Border collies^{12, 48}, the skewed male to females ratio would be perfectly understandable. However, that would not explain the disorders proposed as autosomal recessive with minimal genetic support (Borzoi^{13, 14}), that are sporadic (German shepherd¹⁰, Greyhound¹⁵), or that are associated with OLM (New Zealand working sheepdogs⁹ and Border collies¹¹). An important question to ask is what ‘protects’ females from the disease, or alternatively, ‘favors’ disease expression in males?

Studies have emphasized differences in innate and adaptive immunity between males and females. A mouse model used to test protective immunity against influenza A found that females develop stronger humoral immune responses and greater cross-protection against heterosubtypic viral challenge⁴⁹. C57BL/6 mice show that gonadectomy in both sexes removed the sex difference in influenza A pathogenesis, and intact females administered high doses of estradiol had increased rates of survival after infection⁵⁰. The gender differences are not exclusive to the mouse, and phylogenetically extend from sea urchins to birds, Rhesus macaques, and man, and changes in innate and adaptive immunity occur at different stages of life from in utero/childhood to old age⁵¹. Of note is that male patients have higher hepatitis B or human immunodeficiency virus loads than females, and a male bias in COVID-19 mortality has been observed during the current pandemic⁵². On the basis of the extremely limited information we have on retinopathy/canine DUSN to date, it is difficult to speculate on the reason(s) for this gender imbalance. However, it is important to keep in mind the possibility that gender-specific differences in innate and adaptive immunity may be at play.

Parasite transmission and routes of exposure

Toxocara canis is exquisitely designed to cause LM, including OLM and DUSN, in dogs, and is a highly effective somatic migrator. Dogs become infected with T. canis by ingesting infective eggs from the environment, paratenic hosts, or in utero from hormonally-activated larvae in the tissues of the bitch. Following ingestion of infective eggs, larvae hatch, penetrate the gut and undergo a liver-lung migration; following bronchial and tracheal migration they are swallowed and end up in the gut and mature⁵³. In dogs after ~8 weeks of age, the larvae are considerably more prone to remain in the pulmonary circulation, where they cross to the venous side and are distributed via the systemic circulation. They thus gain access to various tissues, including eye, where they lodge, penetrate and migrate locally, eventually becoming encapsulated as arrested infective larvae⁵³.

Somatic migration is critical for seeding the tissues with encysted larvae. Male dogs represent a dead-end host for the larvae, but in females encysted larvae are reactivated by pregnancy hormones and migrate into the pups in utero, remaining in the lungs until parturition before completing their migration to the intestine. This transplacental transmission is key to the high prevalence of T. canis infection in puppies (many are born with it), and once infected with somatic larvae, a bitch can infect several litters.

Somatic migration of T. canis larvae also occurs in a wide variety of domestic and wild animals, all of which can transmit back to the dog via predation and scavenging⁴⁰, typical biology for most ascarids and some other carnivore nematodes. Less understood is the relative importance of paratenesis, i.e. the reentrance and remigration of larvae in the tissues of the dog when an infected paratenic host is eaten. Due to a differences in/lack of hormonal influences that activate larvae, male dogs might behave more like paratenic hosts, resulting in continued somatic migration. Such sex-related factors affecting larval migration could potentially account for the higher prevalence of DUSN-like OLM in males versus females.

The raccoon ascarid, B. procyonis, is a very strong somatic migrator in >150 species of mammalian and avian paratenic hosts where it causes NLM, OLM, and VLM^{39, 42}. It is a well-recognized cause of human OLM and DUSN, and DUSN lesions were produced experimentally in two species of nonhuman primates and five species of rodents following oral infection with eggs²³. Severe NLM and VLM occurred both naturally and experimentally in dogs, but unfortunately the eyes were not examined³⁹.

Ocular penetration and pathogenesis

Following systemic dissemination, somatically migrating larvae usually gain access to the eye via the ophthalmic artery where they become trapped in small vessels of the choroid, and anatomic considerations dictate the route of entry into the eye. If located in the superior quadrant, it is unlikely that larvae penetrate directly through the tapetal layer as this layer is a thick, multicellular layer, somewhat reminiscent of a brick wall with tightly ordered and compacted tapetal rodlets (Figure 3E–G). However, the tapetum is penetrated by regularly spaced arterioles allowing ready access to the subretinal space (Figure 3A, B). Because of size considerations²⁴, it is likely that the migration is perivascular not intravascular. Regardless, this results in focal infarcts of the tapetal arterioles and the characteristic fundus lesions (Figure 1A, B). In the non-tapetal region, access to the subretinal space is more direct as the highly vascular choroid is juxtaposed to the choriocapillaris (Figure 3C, D). This results in focal depigmented areas that can expand, and often have a telltale central area of focal damage indicating the site of penetration into the retina (e.g. Figures 1E₃ and 2E₂).

Since DUSN is a multi-etiologic OLM syndrome potentially caused by a variety of nematodes (Toxocara, Baylisascaris, Ancylostoma, snake ascarids, etc.) and a trematode (Alaria), its common pathogenesis is likely related to characteristics and/or consequences of migrating larvae. In addition to direct mechanical damage, parasite-induced toxicity in sensitive ocular tissues is likely related to highly antigenic larval excretory-secretory (ES) products as well as toxic proteins released by degranulating/degenerating eosinophils in response to the larvae and their products. Eosinophils accumulate in response to migrating larvae and produce an array of very toxic and damaging substances, particularly affecting the retina^{37, 54}. Retinal migration is accompanied by focal to diffuse disruption and degeneration, hemorrhages, pyknosis of nuclear layer nuclei, hyperplasia and migration of the RPE, and varying degrees of eosinophilic retinitis, choroiditis, vitritis, vasculitis and optic neuritis²³. Developing retinal or choroidal granulomas are often overlain with a vitreal reaction^{37, 54}.

Watzke proposed that T. canis larvae were actively motile and could move freely through ocular tissues, initiating new inflammatory reactions at sites of relocation. The reaction depends on how long a larva remains in a particular location, its ES products and the ensuing immune response⁵⁴. When larvae leave the eye, and in eyes in which larvae have died or been killed, inflammation can subside. However, when significant inflammatory reactions have taken place degenerative and atrophic changes characteristic of DUSN ensue.

If canine DUSN is suspected or diagnosed, what treatments are recommended?

When larvae are visualized in the human eye and are away from critical areas, laser photocoagulation is the preferred treatment^{26, 30, 33, 55}. Because living larvae continue to produce ES antigens and inflammation, it is best to kill them as early and rapidly as possible^{42, 54}. Problem with photocoagulation is finding the larvae in the first place, as this can be quite difficult depending on their particular location^{30, 34}. In human patients, corticosteroids are sometimes suggested to reduce inflammation associated with worm death, but their use is equivocal and they appear unnecessary because photocoagulation destroys the parasite and does not usually exacerbate inflammation. In the T. canis primate model systemic steroids did not affect the inflammation or outcome⁵⁴.

Anthelmintic therapy is an obvious alternative treatment for DUSN, especially in suspected cases where a worm is suspected but not found, or as a stand-alone or additional treatment. Various drugs have been tried in human OLM and DUSN, including thiabendazole, diethylcarbamazine, and ivermectin, with variable and typically unsatisfactory results^{34, 56}. Much better efficacy was seen with high-dose albendazole⁵⁷ which is the drug of choice for DUSN^{30, 31, 34}. Albendazole is a potent broad-spectrum anthelmintic against a variety of helminths in humans and animals, has good tissue distribution, will cross the blood-brain and blood-ocular barriers, and is usually well-tolerated in people^{30, 34, 57, 58}.

But there is a problem with albendazole in dogs as it is not approved for use in nor is it well-tolerated by small animals. Potentially serious adverse reactions in dogs, particularly at high doses, include bone marrow dysplasia, with associated aplastic anemia, leukopenia, thrombocytopenia, and anorexia^{53, 59}. Despite this, it has been used to treat a number of more difficult parasite infections in dogs⁵³. Whether the related benzimidazole, fenbendazole, which is safe and well-tolerated in dogs, or other anthelmintics would be efficacious in the treatment of canine DUSN is unknown.

Apart from photocoagulation of visible larvae, treatment of canine DUSN in general remains challenging. Therefore, it is important to recommend regular (e.g., monthly) parasite treatment and control for dogs, including regular fecal cleanup and decontamination, and to dissuade owners from feeding raw meat diets. This will help decrease the chances of infection through egg ingestion as well as possible exposure to food-borne larvae. It is important to note that common canine anthelmintics kill adult nematodes in the GI tract (not migrating larvae), thus stopping the release of eggs in the feces. They will not prevent nematode eggs from being ingested, hatching, and larvae migrating and causing disease. Currently, the only available preventives for ascarid hatching and migration are pyrantel salts^{39, 53}, but, because they require daily dosing they, fell out of favor with the advent of monthly broad-spectrum “heartworm” products for routine parasite control.

Summary and Conclusions

In this Commentary and Review we have made an effort to bring canine DUSN to the attention of the veterinary ophthalmology community, and highlight how it may be responsible for some of the retinal findings grouped under the retinopathy rubric. It is very unlikely that canine DUSN accounts for all lesions characterized by focal/multifocal tapetal hyperreflectivity with a central ‘pigment’ clump and patchy non-tapetal degeneration. However, these lesions may be a signal that OLM is a possibility, and that preventive measures are indicated. Moreover, when such lesions are associated with retinal granulomas then the diagnostic index should rise from suspicious to probable. After all, dogs live in an environment and ingest materials where chronic and repeated exposure to helminth eggs and larvae is a reality. Taking into account these environmental/behavioral factors is important in separating those retinal abnormalities that are clearly inherited from those that are acquired.